CN113094897B - Method for predicting hardenability of steel for automobile - Google Patents
Method for predicting hardenability of steel for automobile Download PDFInfo
- Publication number
- CN113094897B CN113094897B CN202110371468.1A CN202110371468A CN113094897B CN 113094897 B CN113094897 B CN 113094897B CN 202110371468 A CN202110371468 A CN 202110371468A CN 113094897 B CN113094897 B CN 113094897B
- Authority
- CN
- China
- Prior art keywords
- hardness
- sample
- end quenching
- matrix
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 63
- 239000010959 steel Substances 0.000 title claims abstract description 63
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000010791 quenching Methods 0.000 claims abstract description 147
- 230000000171 quenching effect Effects 0.000 claims abstract description 136
- 239000011159 matrix material Substances 0.000 claims abstract description 91
- 230000002159 abnormal effect Effects 0.000 claims abstract description 17
- 238000002474 experimental method Methods 0.000 claims abstract description 15
- 238000013178 mathematical model Methods 0.000 claims abstract description 8
- 238000012545 processing Methods 0.000 claims abstract description 8
- 239000000126 substance Substances 0.000 claims description 23
- 238000012360 testing method Methods 0.000 claims description 20
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 19
- 238000001816 cooling Methods 0.000 claims description 16
- 239000000203 mixture Substances 0.000 claims description 16
- 230000009466 transformation Effects 0.000 claims description 13
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 claims description 12
- 229910052799 carbon Inorganic materials 0.000 claims description 12
- 238000005507 spraying Methods 0.000 claims description 12
- 238000012549 training Methods 0.000 claims description 11
- 238000002790 cross-validation Methods 0.000 claims description 8
- 238000009826 distribution Methods 0.000 claims description 7
- 238000005259 measurement Methods 0.000 claims description 7
- 239000007921 spray Substances 0.000 claims description 7
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 claims description 3
- 238000005315 distribution function Methods 0.000 claims description 3
- 150000001875 compounds Chemical class 0.000 claims description 2
- 230000001131 transforming effect Effects 0.000 claims description 2
- 230000007246 mechanism Effects 0.000 abstract description 4
- 238000004458 analytical method Methods 0.000 abstract description 3
- 238000004519 manufacturing process Methods 0.000 abstract description 3
- 238000012417 linear regression Methods 0.000 description 4
- 229910052717 sulfur Inorganic materials 0.000 description 4
- 238000013473 artificial intelligence Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 3
- 229910001566 austenite Inorganic materials 0.000 description 3
- 238000009776 industrial production Methods 0.000 description 3
- 239000013585 weight reducing agent Substances 0.000 description 3
- 238000007405 data analysis Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000012067 mathematical method Methods 0.000 description 2
- 238000005272 metallurgy Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 229910052698 phosphorus Inorganic materials 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000004781 supercooling Methods 0.000 description 2
- 238000011282 treatment Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- OAICVXFJPJFONN-UHFFFAOYSA-N Phosphorus Chemical compound [P] OAICVXFJPJFONN-UHFFFAOYSA-N 0.000 description 1
- XUIMIQQOPSSXEZ-UHFFFAOYSA-N Silicon Chemical compound [Si] XUIMIQQOPSSXEZ-UHFFFAOYSA-N 0.000 description 1
- NINIDFKCEFEMDL-UHFFFAOYSA-N Sulfur Chemical compound [S] NINIDFKCEFEMDL-UHFFFAOYSA-N 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 229910045601 alloy Inorganic materials 0.000 description 1
- 239000000956 alloy Substances 0.000 description 1
- 238000005275 alloying Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000011572 manganese Substances 0.000 description 1
- WPBNNNQJVZRUHP-UHFFFAOYSA-L manganese(2+);methyl n-[[2-(methoxycarbonylcarbamothioylamino)phenyl]carbamothioyl]carbamate;n-[2-(sulfidocarbothioylamino)ethyl]carbamodithioate Chemical compound [Mn+2].[S-]C(=S)NCCNC([S-])=S.COC(=O)NC(=S)NC1=CC=CC=C1NC(=S)NC(=O)OC WPBNNNQJVZRUHP-UHFFFAOYSA-L 0.000 description 1
- 239000011574 phosphorus Substances 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 229910052710 silicon Inorganic materials 0.000 description 1
- 239000010703 silicon Substances 0.000 description 1
- 239000011593 sulfur Substances 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/20—Design reuse, reusability analysis or reusability optimisation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Data Mining & Analysis (AREA)
- Pure & Applied Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Analysis (AREA)
- Computing Systems (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Algebra (AREA)
- Geometry (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Investigating And Analyzing Materials By Characteristic Methods (AREA)
Abstract
The invention provides a method for predicting hardenability of steel for an automobile, which comprises the following steps: step 1, performing an automobile steel end quenching experiment, obtaining required data, and constructing a sample matrix; step 2, processing the abnormal data; step 3, establishing an intelligent algorithm prediction model of the starting point of the end quenching curve; step 4, establishing an intelligent algorithm prediction model of the minimum value of the end quenching curve; step 5, establishing an end quenching curve mathematical model; and 6, taking the three models obtained in the steps 3, 4 and 5 as final prediction models, and inputting the steel components and the end quenching distance for the automobile into the final prediction models to obtain the hardenability value. The method adopts the idea of an intelligent algorithm to build the predictive model of the hardenability of the steel, and has higher precision and more close to the actual production value than the model built by the existing pure mechanism analysis method.
Description
Technical Field
The invention relates to a method for predicting hardenability of steel for an automobile.
Background
Along with the deep concept of energy conservation and environmental protection, the automobile weight reduction is more and more widely focused, and the automobile weight reduction can improve the automobile operability, so that the acceleration performance is better, and meanwhile, the energy can be saved. The steel product occupies a large proportion in the total mass of the automobile, and on the premise of ensuring the strength and the safety performance, the toughness of the steel for the automobile is improved as much as possible, and the use amount of the steel is reduced, so that the whole mass of the automobile is lightened, and the steel product is one of the most effective ways for lightening the automobile. The toughness of steel for automobiles mainly depends on the structure state, and the basis for obtaining a certain structure state is a transformation product cooled by high-temperature austenite. The characteristics of the transformation product are directly and closely related to the supercooling capability of austenite, and an important scale for measuring the supercooling capability of austenite is hardenability. Therefore, accurate grasping of the hardenability of each batch of steel is a primary condition for ensuring the balance between the weight reduction and the reliability of the automotive steel product. The end quenching curve obtained by the end quenching experiment can well reflect the performance of the steel, and is the best expression for the steel. Factors influencing the hardenability of steel for automobiles are mainly the degree of alloying, the degree of austenitization and the grain size, and most importantly the chemical composition. Therefore, calculating the predicted end quenching curve using chemical components as basic parameters is an important issue in this field.
At present, researchers at home and abroad conduct intensive research on end quenching curve modeling of steel, and the end quenching curve modeling can be summarized into two types of mechanism modeling and statistical analysis method modeling. The mechanism modeling method is to build a mathematical expression model by analyzing and researching the roles of components, grain size and the like of steel in the hardenability of the steel. However, because of the numerous influencing factors involved in hardenability, the factors are related to each other, and the effects of some factors are not clear, it is difficult to establish an accurate quantitative mathematical expression, and the accuracy of the established model needs to be improved. In the aspect of statistical analysis method modeling, a relevant scholars mainly adopt a regression model and an artificial intelligence algorithm. For example, the application with the Chinese patent application number of 201010287677.X and the application date of 2010-09-19 discloses a hardenability prediction and a method for producing narrow hardenability strip steel, wherein the application combines an artificial neural network model and an incremental algorithm, takes the component increment of chemical components to a reference heat as input, takes the end quenching value increment of Jominy end quenching hardness to the reference heat as output, and establishes a hardenability prediction model based on an incremental neural network; the prior Chinese patent application No. CN201310598566.4 with application date of 2013, 11 and 22 discloses a general relation between the hardenability coefficient and the end hardening degree established by adopting a nonlinear fitting method, and a relation between alloy elements, grain size grade and the hardenability coefficient established by using a support vector machine method. The application with the prior Chinese patent application number of CN201910571054.6 and the application date of 2019, 6 and 28 discloses a hardenability prediction method of artificial neural network steel and a prediction model establishment method thereof. Although the accuracy of the algorithms is high, the predicted value is just the hardness of a plurality of fixed points, and the physical metallurgy meaning is not clear.
The performance of the artificial intelligence algorithm model is highly dependent on data, and is mainly expressed in two aspects, namely the data volume; another aspect is the accuracy and validity of the data. In terms of data volume, the end quenching experiment can be repeated for a plurality of times, and the requirements are easily met. The accuracy and effectiveness of the end quench data, as processed based solely on existing mathematical methods, is inadequate and must be combined with physical metallurgical principles.
Disclosure of Invention
The invention aims to: the invention aims to overcome the defects in the prior art and provides a method for predicting hardenability of steel for an automobile, which comprises the following steps:
step 1, performing an automobile steel end quenching experiment, obtaining required data, and constructing a sample matrix;
step 2, processing the abnormal data;
step 3, establishing an intelligent algorithm prediction model of the starting point of the end quenching curve;
step 4, establishing an intelligent algorithm prediction model of the minimum value of the end quenching curve;
step 5, establishing an end quenching curve mathematical model;
and 6, taking the three models obtained in the steps 3, 4 and 5 as final prediction models, and inputting the steel components and the end quenching distance for the automobile into the final prediction models to obtain the hardenability value.
The step 1 comprises the following steps: according to GB/T225-2006, different batches of different types of automobile steel with known components are manufactured into end quenching samples, then the end quenching samples are heated to austenitizing, then water spray cooling is rapidly carried out on the top ends, after cooling, the water spray ends of the end quenching samples, namely the top ends, are taken as starting points, hardness values are tested at intervals a along the length direction of two sides respectively, until the top ends are reached, and the distance from each test point to the water spray ends, namely the top ends, is called the end quenching distance. After the test is finished, the average value of the two sides is calculated to be used as the hardness value of the end quenching distance, and the hardness values of the end quenching distances are connected to form a curve to obtain an end quenching curve.
Setting the end quenching distance as t, marking the hardness distribution as f (t, C), wherein C is a coefficient related to the chemical composition of the standard sample;
constructing a component matrix X by using chemical components of the end-quenched sample, wherein the element X in the component matrix X i,j A j-th component of the i-th end-quenched sample;
constructing a hardness matrix H by using the hardness values of all test points of the end quenching sample, wherein the element H in the hardness matrix H i,j A hardness value of a j-th measurement point of the i-th end quenching sample;
the composition and hardness distribution of each end-quenched sample is referred to as one sample, and the aggregate of all samples is referred to as the sample matrix.
The step 2 comprises the following steps:
step 2-1, eliminating abnormal values according to the relation between the carbon content and the hardness;
step 2-2, eliminating abnormal values according to the relation between the cooling rate and the hardness;
step 2-3, eliminating abnormal values according to component content standards;
and 2-4, performing rechecking inspection on the sample matrix.
Step 2-1 includes: arranging samples according to the carbon content in the component matrix X from small to large, marking the sample of the front b as Xmax, marking the sample of the rear c as Xmin, and then marking the sample as the hardness value M of the initial point of the end quenching curve according to the first column of the matrix H i,1 Arranging from small to large, marking the sample of the front d as Hmax, marking the sample of the rear e as Hmin, and performing a collective operation, such as the formulas (1), (2):
A 1 ＝Xmax∩Hmin (1)
A 2 ＝Xmin∩Hmax (2)
obtaining intersection A 1 And A 2 ；
The starting point of the end quenching curve is the position closest to the water spraying end on the end quenching sample, and the conditions should be satisfied: high carbon content and high hardness, deleting intersection A 1 And A 2 Not satisfying the condition.
Step 2-2 includes: in the end quenching experiment, the cooling rate of the end quenching sample is the largest at the water spraying end, the hardness is the highest, then the cooling rate is gradually reduced along with the increase of the end quenching distance, and the hardness and the cooling rate are directly related and also sequentially reduced. Based on the basic theory, the treatment is carried out by adopting the steps from 2-2-1 to 2-2-3:
step 2-2-1, a transformation matrix A is established, as shown in a formula (3):
the hardness matrix H is transformed, and a new matrix is constructed and marked as M, as shown in a formula (4):
M＝HA (4)
the relation between the matrix M obtained after transformation and the H matrix before transformation is shown in the formula (5):
m i,1 ＝h i,1 ，m i,j ＝h i,j-1 -h i,j (5)
in formula (5), h i,j Represents the hardness value, m, of the j-th measurement point of the i-th end quenching sample in the hardness matrix H i,j Is the hardness value of the previous test point of the i-th end quenching sample minus the hardness value of the next test point.
Step 2-2-2, if m i,j Less than or equal to 0, the hardness of the j-th measuring point of the i-th sample is not reduced along with the distance from the water spraying end, and the i-th sample is removed from the sample matrix;
and 2-2-3, after the sample matrix is removed, reversely reconstructing the sample matrix by using the inverse matrix of the transformation matrix A.
The step 2-3 comprises the following steps:
each column of the component matrix X is a certain component content of the automobile steel, and samples exceeding the standard are removed according to the technical standard of the automobile steel.
The step 3 comprises the following steps:
the starting point of the end quenching curve is also the maximum value of the end quenching curve, and is marked as J max (the first intelligent model is J max It is directly applied in equation (6), defining the parameters of the functional formula K, k=jmax);
taking the component matrix X as input, taking the first column of the hardness matrix H, namely the set of hardness values of the initial points of the sample end quenching curves as output, and constructing a data set S 0 ；
Training an artificial intelligent algorithm model by using a K-fold cross validation method, and collecting a data set S 0 Dividing the model into K equal parts, taking K-1 data in the model as training data, taking the left part as test data, respectively carrying out K iterative operations on more than two artificial intelligent algorithm models, calculating an accuracy average value, and selecting the artificial intelligent algorithm model with the highest average value as an end quenching curve starting point intelligent algorithm prediction model.
Step 4 comprises:
the end hardness value of the end quenching sample is the end quenching curve minimum value, which is recorded as J min (the second intelligent model is J min It is directly applied to the formula (6), J min 、J max Together determining parameter a);
taking the component matrix X as input, and taking the last column of the hardness matrix H, namely the end hardness value set of the sample end quenching curve as output, to construct a data set S n ；
Training an artificial intelligent algorithm model by using a K-fold cross validation method, and collecting a data set S n Dividing the model into K equal parts, taking K-1 data as training data, taking the left part as test data, respectively carrying out K iterative operations on more than two artificial intelligent algorithm models, calculating an accuracy average value, and selecting the artificial intelligent algorithm model with the highest average value as an end quenching curve minimum intelligent algorithm prediction model.
The step 5 comprises the following steps:
step 5-1, setting the end quenching distance as t, recording the hardness distribution function as J (t), and constructing a fitting function J of an end quenching curve, wherein the fitting function J is shown in a formula (6):
wherein t represents the end quenching distance; k and a are undetermined coefficients, and r is a chemical component parameter; f (t, C) above indicates that hardness is a function of end quench distance and composition. More of J (t) herein only considers hardness as a function of end quench distance;
step 5-2, obtaining the maximum J of hardness when t=0 max The value is obtained by the model established in the step 3;
when t.fwdarw.0, the compound is obtained by the formula (6):
obtaining, K=J max ；
In the step 5-3 of the method, when t → + in the case of infinity, the air conditioner is controlled, obtaining the minimum value J of the hardness min The value is obtained by modeling in the step 4;
when the temperature is t to + & gt, obtained by (5):
because K=J max Obtaining
Step 5-4, obtaining the formula (7) by transforming the formula (6)
Setting intermediate parametersIntermediate parameters->Obtaining the formula (8):
r is a function of the composition content of the automotive steel, as shown in formula (9):
r＝a 0 +a 1 *[C]+a 2 *[M n ]+a 3 *[S i ]+a 4 *[C r ]+a 5 *[N i ]+a 6 *[S]+a 7 *[M o ]+a 8 *[B] (9)
in the formula (9), a 0 ，a 1 ，a 2 ，a 3 ，a 4 ，a 5 ，a 6 ，a 7 ，a 8 For undetermined coefficients, [ C ]]，[M n ]，[S i ]，[C r ]，[N i ]，[S]，[M o ]，[B]Respectively C, M n ，S i ，C r ，N i ，S，M o The mass percentage of the chemical components B in the end quenching sample;
combining formula (8) and formula (9) to obtain formula (10):
taking a component matrix X as input, taking a hardness matrix H and end quenching distance as output, and solving a through an artificial intelligent regression algorithm 0 ，a 1 ，a 2 ，a 3 ，a 4 ，a 5 ，a 6 ，a 7 ，a 8 And then obtaining the mathematical model of the end quenching curve.
The main innovation points of the invention are as follows:
(1) The method combining the mathematical method and the physical metallurgy principle is adopted, and a processing method of the steel hardenability data for the automobile is innovated.
(2) And constructing a new hardenability prediction mathematical model by utilizing an artificial intelligence algorithm according to the steel hardenability distribution for the automobile.
The beneficial effects are that: the method adopts the idea of an intelligent algorithm to build the predictive model of the hardenability of the steel, and has higher precision and more close to the actual production value than the model built by the existing pure mechanism analysis method. The predicted value of the built model can completely or partially replace a physical experiment value, so that a great amount of manpower is saved, and pollution and energy consumption are reduced. The model can provide reference for the design of new steel components, improve the development success rate of the new steel, and reduce development cost and period.
Drawings
The foregoing and/or other advantages of the invention will become more apparent from the following detailed description of the invention when taken in conjunction with the accompanying drawings and detailed description.
FIG. 1 is a flow chart of predicting the end quenching curve of the steel for the automobile.
FIG. 2 is a schematic diagram of an abnormal data analysis, culling and processing flow.
Fig. 3 is a schematic diagram of a top quench and end quench curve test.
Detailed Description
As shown in fig. 1 and 2, the invention provides a method for predicting hardenability of steel for an automobile, comprising the following steps:
step 1, performing an automobile steel end quenching experiment to obtain required data and constructing a data set;
step 2, analyzing, rejecting and processing abnormal data;
step 3, predicting a model by an intelligent algorithm of the starting point of the end quenching curve;
step 4, predicting a model by an intelligent algorithm of the minimum value of the end quenching curve;
step 5, establishing an end quenching curve mathematical model;
and 6, taking the three models obtained in the steps 3, 4 and 5 as final prediction models. And inputting the steel components and the end quenching distance into a model to obtain the hardenability value.
The step 1 comprises the following steps: according to GB/T225-2006, standard samples with diameters of 24mm and lengths of 100mm are prepared from different batches of different types of automobile steel with known components, then the standard samples are heated to austenitization, water spray cooling is rapidly carried out on the tail ends, after cooling, hardness values are measured at intervals a (generally 1 mm) along the length directions of two sides by taking the top ends of quenching ends of the samples as starting points, and the hardness values are measured until the tail ends. After the test is finished, the average value of the two sides is calculated as the hardness value of the distance, the distance from the measuring point to the water spraying end is called as the end quenching distance, and the hardness values on the end quenching distances are connected to form a curve, namely the end quenching curve, as shown in fig. 3.
Let the end quenching distance be t, the hardness distribution is directly related to the end quenching distance and the sample components, denoted as f (t, C), t is the end quenching distance, and C is the coefficient related to the components.
The components of the end-quenched test sample are as follows: C. mn, si, cr, ni, S, mo, B A matrix of components is constructed, denoted as X, with the elements X in the matrix i,j I represents the ith sample, j represents the jth element.
Constructing a hardness matrix by using the hardness values of all test points of the end quenching experimental sample, and marking the hardness matrix as H, wherein the element H in the matrix i,j I represents the i-th sample, and j represents the j-th measurement point.
The composition and hardness value distribution of each sample is collectively referred to as a sample. The aggregate of all samples is called the sample matrix, which obviously consists of the component matrix X and the stiffness matrix H.
The step 2 comprises the following steps:
step 2-1, eliminating abnormal values according to the relation between the carbon content and the hardness
Arranging samples according to the first column of the input matrix X, namely the carbon content, from small to large, marking the sample of the front b (the value is 10 percent generally) as Xmax, marking the sample of the rear c (the value is 10 percent generally) as Xmin, and then marking the sample according to the hardness value M of the first column of the matrix M, namely the starting point of the end quenching curve i,1 Arranging from small to large, marking the sample of the front d (with the value of 10 percent generally) as Mmax, marking the sample of the rear e (with the value of 10 percent generally) as Mmin, and performing the aggregation operation, such as the formulas (1), (2):
A 1 ＝Xmax∩Mmin (1)
A 2 ＝Xmin∩Mmax (2)
obtaining intersection A 1 And A 2 。
The starting point of the end quenching curve is the position closest to the water spraying end on the sample, the hardness value of the end quenching curve mainly depends on the carbon content, and the hardness is high when the carbon content is high. Intersection A 1 And A 2 The samples contained in the sample are consistent with the principleViolation should be eliminated; step 2-2, eliminating abnormal values according to the relation between the cooling rate and the hardness
In the end quenching experiment, the cooling rate of the sample at the water spraying end is the largest, the hardness is the highest, then the cooling rate is gradually reduced along with the increase of the end quenching distance, and the hardness is also sequentially reduced due to the direct correlation of the hardness and the cooling rate. Based on this basic theory, the treatments were performed in the modes (1) to (3).
(1) And (3) performing a transformation matrix A, as shown in a formula (3):
the hardness matrix H is transformed, and a new matrix is constructed and marked as M, as shown in a formula (4):
M＝HA (4)
the relation between the matrix M obtained after transformation and the H matrix before transformation is shown in the formula (5):
m i,1 ＝h i,1 ，m i,j ＝h i,j-1 -h i,j (5)
i represents a matrix row index, j represents a matrix column index, and j is ≡2.
(2) If matrix H i,j Less than or equal to 0, that is, the hardness of the jth measurement point of the ith sample does not decrease with distance from the water spraying end, which is obviously contrary to the basic principle that the sample should be removed from the sample matrix;
(3) And after the sample matrix is removed, reversely reconstructing the sample matrix by using the inverse matrix of the matrix A.
Step 2-3, eliminating abnormal values according to the component content standard
Each column of the component matrix X is a certain component content of the automobile steel, and samples exceeding the standard are removed according to the technical standard of the automobile steel. As shown in table 1, samples that exceeded this standard were rejected as car steel 510L.
TABLE 1 automobile Steel 510L chemical composition (mass percent)
In Table C, S i ，M n P and S represent chemical components such as carbon, silicon, manganese, phosphorus, sulfur and the like in steel.
And 2-4, rechecking and checking the removed sample data, namely, re-carrying out end quenching experiments on the samples corresponding to the data, if the end quenching experiments are caused by experimental errors, inputting new experimental result data into a matrix, and re-carrying out analysis and removal of the sample matrix, and if the end quenching experiments are not caused by experimental errors, analyzing the reasons caused by research.
The step 3 comprises the following steps:
the initial point of the end quenching curve is the position closest to the water spraying end on the sample, the point is the point with the highest quenching degree on the sample, the point is the maximum value of the end quenching curve, and the point value is marked as J max 。
Taking the component matrix X as input, taking the first column of the hardness matrix H, namely the set of hardness values of the initial point of the sample end quenching curve as output, and constructing a data set S 0 。
Training an artificial intelligent algorithm model by using a K-fold cross validation method, and collecting a data set S 0 Dividing the model into K equal parts, taking K-1 data as training data, taking the left part as test data, respectively carrying out K iterative operations on more than two artificial intelligent algorithm models, calculating an accuracy average value, and selecting the artificial intelligent algorithm model with the highest average value as a prediction model of the end quenching curve initial point hardness value.
Step 4 comprises:
the minimum value of the end quenching curve is theoretically infinity from the water spraying end, the hardness value is not influenced by quenching, is approximate to the normalized hardness, can be approximately replaced by the hardness value of the end of the sample, and is recorded as J min 。
The composition matrix X is taken as input, and the last column of the hardness matrix H, namely the hardness value set of the end of the sample end quenching curve is taken as output to construct a data set S n 。
Method for performing K-fold cross validation on artificial intelligent algorithm modelTraining, data set S n Dividing the model into K equal parts, taking K-1 data as training data, taking the left part as test data, respectively carrying out K iterative operations on more than two artificial intelligent algorithm models, calculating an accuracy average value, and selecting the artificial intelligent algorithm model with the highest average value as a prediction model of the end quenching curve end hardness value.
The step 5 comprises the following steps:
5-1: and (3) setting the end quenching distance as t, recording the hardness distribution function as J (t), and constructing a fitting function of an end quenching curve as shown in a formula (6).
In the formula (6), t represents a terminal quenching distance; k, a is a coefficient to be determined; r is a chemical component parameter.
5-2: when t=0, the maximum hardness J is obtained max This value is obtained from the model in step (3).
When t.fwdarw.0, formula (6) can be obtained:
obtaining, K=J max 。
5-3: when t → + in the case of infinity, the air conditioner is controlled, obtaining the minimum value J of the hardness min This value is obtained from the model in step (4).
When t.fwdarw++ is present, the formula (6) can be obtained:
because K=J max Can be obtained by
5-4: the formula (6) is converted into the formula (7)
Is provided with
Formula (8) is available:
r is a function of the content of the steel components of the automobile, as shown in formula (9)
r＝a 0 +a 1 *[C]+a 2 *[M n ]+a 3 *[S i ]+a 4 *[C r ]+a 5 *[N i ]+a 6 *[S]+a 7 *[M o ]+a 8 *[B] (9)
In the formula (9), a 0 ，a 1 ，a 2 ，a 3 ，a 4 ，a 5 ，a 6 ，a 7 ，a 8 For undetermined coefficients, [ C ]]，[M n ]，[S i ]，[C r ]，[N i ]，[S]，[M o ]，[B]C, M respectively n ，S i ，C r ，N i ，S，M o The mass percentage of the chemical component B in the end quenching sample.
Combining formula (8) and formula (9) gives formula (10):
a is obtained by taking a component matrix X as input, taking a hardness matrix H and an end quenching distance as output and an artificial intelligent regression algorithm 0 ，a 1 ，a 2 ，a 3 ，a 4 ，a 5 ，a 6 ，a 7 And further obtaining a predictive model.
Examples 1, 35MnB end quench Curve prediction
(1) Performing an automobile steel end quenching experiment to obtain required data and constructing a data set
End quenching experiments were performed on automotive steel 35MnB, and the cumulative sample data totaled 625 groups.
(2) Abnormal data analysis, rejection and processing
2.1 reject outlier 35 samples according to carbon content versus hardness.
2.2 eliminating 15 groups of samples with abnormal values according to the relation between the cooling rate and the hardness
2.3 eliminating abnormal values according to the component content standard
The ranges of the main elements are set according to the steel technology standard for automobiles, as shown in table 2.
TABLE 2 automobile Steel 35MnB chemical composition (mass fraction)
And eliminating 2 out of standard samples after inspection.
2.4, after re-experiments on the samples from which the samples are removed, the experimental result data change is small, so that the sample matrix should not be re-entered.
Step 3, an intelligent algorithm prediction model for starting point of end quenching curve
The first column of the sample stiffness matrix, i.e., the set of sample end quench curve start end stiffness values, is used as an input to construct a data set.
After normalization processing is carried out on the data set, an artificial intelligent algorithm model is selected by adopting a K-fold cross validation method. The k=6 verification method is adopted to compare the model LinearRegression, LASSO, elasticNet, KNeighborsRegressor, decisionTreeRegressor, SVR and the like, and the average value of the accuracy is as follows:
LinearRegression：0.657393
LASSO：0.697506
ElasticNet：0.697506
KNeighborsRegressor：0.853791
DecisionTreeRegressor：0.963117
SVR：0.712025
therefore, the DecisionTreeRegresor model is selected as a predictive model of the starting point of the end quench curve, i.e., the maximum of the end quench curve.
Step 4, end quenching curve minimum intelligent algorithm prediction model
The minimum value of the end quenching curve is theoretically infinity from the water spraying end, the hardness value is not influenced by quenching, is approximate to the normalized hardness, can be approximately replaced by the hardness value of the end of the sample, and is recorded as J min 。
The sample component matrix is used as input, and the last column of the sample hardness matrix, namely the end hardness value set of the sample end quenching curve, is used as output to construct a data set.
After normalization processing is carried out on the data set, an artificial intelligent algorithm model is selected by adopting a K-fold cross validation method. The k=6 verification method is adopted to compare the model LinearRegression, LASSO, elasticNet, KNeighborsRegressor, decisionTreeRegressor, SVR and the like, and the average value of the accuracy is as follows:
LinearRegression：0.670783
LASSO：0.690237
ElasticNet：0.790327
KNeighborsRegressor：0.452148
DecisionTreeRegressor：0.934266
SVR：0.7 11378
therefore, the DecisionTreeRegresor model was chosen as the predictive model for the end-quench curve minima.
Step 5, establishing an end quenching curve mathematical model
End quenching curve function expressionThe parameter K of (a) is obtained by step 3 and a is obtained by step 4.
Through an artificial intelligent regression algorithm, the following steps are obtained:
r＝1.3+63.9*[C]+12*[Mn]+3.0*[Si]+13.6*[Cr]+3.3*[Ni]-4*[S]+28*[Mo]-3*[B]
in the formula, [ C ]]，[M n ]，[S i ]，[C r ]，[N i ]，[S]，[M o ]，[B]C, M respectively n ，S i ，C r ，N i ，S，M o The mass percentage of the chemical composition B in the steel grade of the automobile to be predicted.
And 6, taking the three models obtained in the steps 3, 4 and 5 as final prediction models. And inputting the steel composition and the end quenching distance of the automobile steel into a model to obtain the hardenability value. Table 4 shows the comparison between the experimental values and the predicted values of the samples of the components shown in Table 3.
Automobile steel 35MnB chemical composition (mass fraction) of Table 3, sample No. 16711421
Table 4, experimental and predictive value comparisons
It can be seen from table 4 that the maximum error point is at the position of 15mm in the end quenching distance, and the maximum error percentage is 5.2%, so that the precision requirement required by industrial production can be met.
Example 2, 42CrMoS4HH end quench Curve prediction
Automobile steel 42CrMoS4HH chemical composition (mass fraction) of Table 5, sample No. 14708767
Table 6, comparison of experimental and predicted values
| End quenching distance (mm) | 1.5 | 5 | 9 | 13 | 15 | 20 | 25 | 30 | 35 | 40 | 45 | 50 |
| Actual measurement value | 59.4 | 58 | 56.5 | 54.6 | 52.6 | 47.2 | 44.35 | 42.35 | 40.45 | 38.2 | 37.45 | 35.5 |
| Predictive value | 59.8 | 59 | 57.61 | 53.79 | 51.79 | 47.4 | 44 | 41.44 | 39.47 | 37.93 | 36.7 | 35.7 |
It can be seen from table 6 that the maximum error point is at the position of 9 end quenching distance, and the maximum error percentage is 2%, so that the precision requirement required by industrial production can be met.
Example 3, 1E1287 end quenching Curve prediction
Automobile steel 1E1287 chemical components (mass fraction) of Table 7, sample No. 16711348
Table 8, comparison of experimental and predicted values
It can be seen from table 8 that the maximum error point is at the position of 7mm in the end quenching distance, and the maximum error percentage is 6.4%, so that the precision requirement required by industrial production can be met.
The invention provides a method for predicting hardenability of steel for an automobile, and the method and the way for realizing the technical scheme are numerous, the above description is only a preferred embodiment of the invention, and it should be pointed out that, for a person skilled in the art, several improvements and modifications can be made without departing from the principle of the invention, and the improvements and modifications should also be regarded as the protection scope of the invention. The components not explicitly described in this embodiment can be implemented by using the prior art.
Claims (1)
1. The method for predicting the hardenability of the steel for the automobile is characterized by comprising the following steps of:
step 1, performing an automobile steel end quenching experiment, obtaining required data, and constructing a sample matrix;
step 2, processing the abnormal data;
step 3, establishing an intelligent algorithm prediction model of the starting point of the end quenching curve;
step 4, establishing an intelligent algorithm prediction model of the minimum value of the end quenching curve;
step 5, establishing an end quenching curve mathematical model;
step 6, taking the three models obtained in the step 3, the step 4 and the step 5 as final prediction models, and inputting the steel components and the end quenching distance for the automobile into the final prediction models to obtain the hardenability value;
the step 1 comprises the following steps: according to GB/T225-2006, different batches of different types of automobile steel with known components are manufactured into end quenching samples, then the end quenching samples are heated to austenitizing, then water spray cooling is rapidly carried out on the top ends, after cooling, the water spray ends of the end quenching samples, namely the top ends, are used as starting points, each hardness value is tested at intervals a along the length direction of two sides until reaching the tail ends, and the distance from each test point to the water spray ends, namely the tail ends, is called end quenching distance; after the test is finished, the average value of the two sides is calculated to be used as the hardness value of the end quenching distance, and the hardness values of the end quenching distances are connected to form a curve to obtain an end quenching curve;
setting the end quenching distance as t, marking the hardness distribution as f (t, C), wherein C is a coefficient related to the chemical composition of the end quenching sample;
constructing a component matrix X by using chemical components of the end-quenched sample, wherein the element X in the component matrix X i,j A j-th chemical component representing an i-th end-quenched sample;
constructing a hardness matrix H by using the hardness values of all test points of the end quenching sample, wherein the element H in the hardness matrix H i,j A hardness value of a j-th measurement point of the i-th end quenching sample;
the chemical composition and hardness distribution of each end-quenched sample is referred to as one sample, and the aggregate of all samples is referred to as a sample matrix;
the step 2 comprises the following steps:
step 2-1, eliminating abnormal values according to the relation between the carbon content and the hardness;
step 2-2, eliminating abnormal values according to the relation between the cooling rate and the hardness;
step 2-3, eliminating abnormal values according to component content standards;
step 2-4, rechecking and checking the sample matrix;
step 2-1 includes: arranging samples according to the content of carbon components in a component matrix X from small to large, marking the sample of the front b as Hmax and the sample of the rear c as Xmin, and then marking the sample as the hardness value H of the initial point of an end quenching curve according to the first column of a hardness matrix H i,1 Arranging from small to large, marking the sample of the front d as Hmax, marking the sample of the rear e as Hmin, and performing a collective operation, such as the formulas (1), (2):
A 1 ＝Xmax∩Hmin (1)
A 2 ＝Xmin∩Hmax (2)
obtaining intersection A 1 And A 2 ；
The starting point of the end quenching curve is the position closest to the water spraying end on the end quenching sample, and the conditions should be satisfied: high carbon content and high hardness, deleting intersection A 1 And A 2 Not satisfying the condition;
step 2-2 includes:
step 2-2-1, a transformation matrix A is established, as shown in a formula (3):
the hardness matrix H is transformed, and a new matrix is constructed and marked as M, as shown in a formula (4):
M＝HA (4)
the relation between the matrix M obtained after transformation and the H matrix before transformation is shown in the formula (5):
m i,1 ＝h i,1 ，m i,j ＝h i,j-1 -h i,j (5)
in the formula (5), h i,j Represents the hardness value, m, of the j-th measurement point of the i-th end quenching sample in the hardness matrix H i,h The hardness value of the previous test point of the i-th end quenching sample is subtracted by the hardness value of the next test point;
step 2-2-2, if m i,j If the hardness of the j-th measuring point of the i-th sample is smaller than or equal to 0, the hardness of the j-th measuring point of the i-th sample is not reduced along with the distance from the water spraying end, and the i-th sample is removed from the sample matrix;
step 2-2-3, after eliminating the sample matrix, reversely reconstructing the sample matrix by using the inverse matrix of the transformation matrix A;
the step 2-3 comprises the following steps:
each column of the component matrix X is the chemical component content of the steel for the automobile, and samples exceeding the standard are removed according to the technical standard of the steel for the automobile;
the step 3 comprises the following steps:
the starting point of the end quenching curve is also the maximum value of the end quenching curve, and is marked as J max ；
Taking the component matrix X as input, taking the first column of the hardness matrix H, namely the set of hardness values of the initial points of the sample end quenching curves as output, and constructing a data set S 0 ；
Training an artificial intelligent algorithm model by using a K-fold cross validation method, and collecting a data set S 0 Dividing into K equal parts, using K-1 data in the K equal parts as training data and one part as test data, respectively performing K iterative operations on more than two artificial intelligent algorithm models, calculating an accuracy average value, and selecting the artificial intelligent algorithm with the highest average valueThe model is used as an intelligent algorithm prediction model of the starting point of the end quenching curve;
step 4 comprises:
the end hardness value of the end quenching sample is the end quenching curve minimum value, which is recorded as J min ；
Taking the component matrix X as input, and taking the last column of the hardness matrix H, namely the end hardness value set of the sample end quenching curve as output, to construct a data set S n ；
Training an artificial intelligent algorithm model by using a K-fold cross validation method, and collecting a data set S n Dividing the model into K equal parts, taking K-1 data in the model as training data, taking the left part as test data, respectively carrying out K iterative operations on more than two artificial intelligent algorithm models, calculating an accuracy average value, and selecting the artificial intelligent algorithm model with the highest average value as an end quenching curve minimum intelligent algorithm prediction model;
the step 5 comprises the following steps:
step 5-1, setting the end quenching distance as t, recording the hardness distribution function as J (t), and constructing a fitting function J of an end quenching curve, wherein the fitting function J is shown in a formula (6):
wherein t represents the end quenching distance; k and a are undetermined coefficients, and r is a chemical component parameter;
step 5-2, obtaining the maximum J of hardness when t=0 max The value is obtained by the model established in the step 3;
when t.fwdarw.0, the compound is obtained by the formula (6):
obtaining, K=J max ；
In the step 5-3 of the method, when t → + in the case of infinity, the air conditioner is controlled, obtaining the minimum value J of the hardness min The value is obtained by modeling in the step 4;
when the temperature is t to + & gt, obtained by (5):
because K=J max Obtaining
Step 5-4, obtaining the formula (7) by transforming the formula (6)
Setting intermediate parametersIntermediate parameters->Obtaining the formula (8):
r is a function of the composition content of the automotive steel, as shown in formula (9):
r＝a 0 +a 1 *[C]+a 2 *[M n ]+a 3 *[S i ]+a 4 *[N r ]+a 5 *[N i ]+a 6 *[S]+a 7 *[M o ]+a 8 *[B] (9)
in the formula (9), a 0 ，a 1 ，a 2 ，a 3 ，a 4 ，a 5 ，a 6 ，a 7 ，a 8 For undetermined coefficients, [ C ]]，[M n ]，[S i ]，[C r ]，[N i ]，[S]，[M o ]，[B]Respectively C, M n ，S i ，C r ，N i ，S，M o The mass percentage of the chemical components B in the end quenching sample;
combining formula (8) and formula (9) to obtain formula (10):
taking a component matrix X as input, taking a hardness matrix H and end quenching distance as output, and solving a through an artificial intelligent regression algorithm 0 ，a 1 ，a 2 ，a 3 ，a 4 ，a 5 ，a 6 ，a 7 ，a 8 And then obtaining the mathematical model of the end quenching curve.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110371468.1A CN113094897B (en) | 2021-04-07 | 2021-04-07 | Method for predicting hardenability of steel for automobile |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110371468.1A CN113094897B (en) | 2021-04-07 | 2021-04-07 | Method for predicting hardenability of steel for automobile |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113094897A CN113094897A (en) | 2021-07-09 |
| CN113094897B true CN113094897B (en) | 2023-10-20 |
Family
ID=76674423
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110371468.1A Active CN113094897B (en) | 2021-04-07 | 2021-04-07 | Method for predicting hardenability of steel for automobile |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113094897B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102033978A (en) * | 2010-09-19 | 2011-04-27 | 首钢总公司 | Method for forecasting and producing narrow hardenability strip steel by hardenability |
| CN110287451A (en) * | 2019-06-28 | 2019-09-27 | 安徽工业大学 | A kind of the prediction of hardenability method and its prediction model method for building up of artificial neural network steel |
-
2021
- 2021-04-07 CN CN202110371468.1A patent/CN113094897B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102033978A (en) * | 2010-09-19 | 2011-04-27 | 首钢总公司 | Method for forecasting and producing narrow hardenability strip steel by hardenability |
| CN110287451A (en) * | 2019-06-28 | 2019-09-27 | 安徽工业大学 | A kind of the prediction of hardenability method and its prediction model method for building up of artificial neural network steel |
Non-Patent Citations (2)
| Title |
|---|
| 基于BP神经网络的CrMnBH类钢淬透性预报;李长宏;王萍;沈千成;黄贞益;吴勇;;热加工工艺(第20期);全文 * |
| 钢的淬透性预测模型研究进展;张国强;王毛球;曹燕光;陈思联;;金属热处理(第04期);第224-227页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113094897A (en) | 2021-07-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114611844B (en) | Method and system for determining alloy addition amount in converter tapping process | |
| Qiao et al. | Application of improved GRNN model to predict interlamellar spacing and mechanical properties of hypereutectoid steel | |
| CN103761423B (en) | A kind of hot rolled plate microstructure and property prediction method based on PSO-ELM | |
| US20180260717A1 (en) | Microalloyed steel mechanical property prediction method based on globally additive model | |
| CN114897227B (en) | Multi-steel mechanical property forecasting method based on improved random forest algorithm | |
| Wu et al. | Prediction of the Charpy V-notch impact energy of low carbon steel using a shallow neural network and deep learning | |
| CN112100745B (en) | Automobile girder steel mechanical property prediction method based on LDA theory | |
| CN101927267B (en) | Control method and device for cleaning between rolls of finish rolling strip steel | |
| CN112949005B (en) | High-strength and high-toughness steel design method based on machine learning guidance | |
| CN110802115B (en) | Data-driven hot continuous rolling finish rolling inlet temperature prediction method | |
| CN103886371A (en) | Method for controlling component and thermal treatment technological process of pre-hardening plastic die steel | |
| CN105363794A (en) | Finish rolling energy-saving control method based on mechanical performance prediction model and rolling energy consumption model | |
| CN101320031A (en) | Austenitic stainless steel accurate steel strip performance prediction model and its cold rolling process planning | |
| CN117494531B (en) | Medium carbon steel decarburization depth prediction method based on finite element and XGBoost algorithm | |
| CN113987949A (en) | Plate strip steel deformation resistance prediction method based on data driving | |
| CN113094897B (en) | Method for predicting hardenability of steel for automobile | |
| EP3913635A1 (en) | Metal material design support method and design support device | |
| Wang et al. | Modeling the quantitative effect of alloying elements on the Ms temperature of high carbon steel by artificial neural networks | |
| CN116825253B (en) | Method for establishing hot rolled strip steel mechanical property prediction model based on feature selection | |
| CN111933221B (en) | Method for predicting dynamic recrystallization fraction of Nb microalloyed steel | |
| CN102286655A (en) | Device and method for isothermal normalizing utilizing forging waste heat | |
| CN103678921B (en) | A kind of analysis method of the texture influence non-orientation silicon steel magnetic property based on principle component regression | |
| Dong et al. | Just-in-time learning-based soft sensor for mechanical properties of strip steel via multi-block weighted semisupervised models | |
| Trzaska | Neural networks model for prediction of the hardness of steels cooled from the austenitizing temperature | |
| Zhikharev et al. | Artificial Intelligence and Machine Learning In Metallurgy. Part 2. Application Examples |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |